Hand Pose Detection Using Palm Surface Distance Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Developers of artificial reality applications face challenges in accurately tracking and differentiating between various hand poses, particularly in distinguishing between pinch grabs and whole-hand grabs, which is labor-intensive and difficult to implement across a wide range of applications.
Innovation Solution
A method is provided to continuously detect and monitor hand poses by analyzing images of a user's hand, using techniques such as determining pinch and whole-hand grab strengths based on finger positions and distances, allowing for seamless interaction with objects in artificial reality environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional hand pose tracking methods are used, then basic hand position detection is achievable, but accurate differentiation between pinch grabs and whole-hand grabs is labor-intensive and difficult to implement
Solution Approach 1:
The hand is segmented into multiple detectable portions (fingers, palm, thumb) with specific pose values assigned to each. This segmentation allows the system to analyze individual finger positions and combinations to differentiate between pinch grabs (specific finger combinations) and whole-hand grabs (all fingers curled), providing accurate classification without complex overall hand analysis
Solution Approach 2:
The system uses parameter changes by monitoring pose values of hand portions over time and comparing them against threshold values. By tracking changes in finger positions and calculating grab strength parameters, the system can distinguish between different grab types through quantitative parameter analysis rather than qualitative assessment
2Measurement precision
If comprehensive hand pose analysis is implemented to distinguish all grab types, then interaction accuracy with various objects improves, but processing time and computational resources increase
Solution Approach 1:
The system applies partial action by focusing detection efforts on critical hand portions and pose values relevant to grab detection. Rather than analyzing every aspect of hand motion, the system monitors specific finger positions and pose thresholds that are sufficient to differentiate grab types, reducing computational overhead while maintaining accuracy
Solution Approach 2:
The system continuously monitors hand pose parameters and provides feedback by comparing current pose values against threshold values. This feedback mechanism allows for real-time grab state determination through iterative comparison and adjustment, enabling fast response times while maintaining detection accuracy through continuous refinement
Data Source
AI summary
A method includes a computing system receiving an image of a real-world environment, the image including at least a portion of a hand of a user of an artificial reality device, the hand comprising a palm and a plurality of fingers. The computing system determines a hand pose of the hand using the image, and defines, based on the hand pose, a three-dimensional surface positioned in the palm of the hand. The computing system determines, based on the hand pose, distances between predetermined portions of the plurality of fingers and the three-dimensional surface. The computing system assigns, based on the distances, a pose value for each of the plurality of fingers of the hand and determines, based on the pose values for the plurality of fingers, a grab state of the hand.